spi: add multi-lane support

Merge series from David Lechner <dlechner@baylibre.com>:

This series is adding support for SPI controllers and peripherals that
have multiple SPI data lanes (data lanes being independent sets of
SDI/SDO lines, each with their own serializer/deserializer).

This series covers this specific use case:

+--------------+    +---------+
| SPI          |    | SPI     |
| Controller   |    | ADC     |
|              |    |         |
|          CS0 |--->| CS      |
|         SCLK |--->| SCLK    |
|          SDO |--->| SDI     |
|         SDI0 |<---| SDOA    |
|         SDI1 |<---| SDOB    |
|         SDI2 |<---| SDOC    |
|         SDI3 |<---| SDOD    |
+--------------+     +--------+

The ADC is a simultaneous sampling ADC that can convert 4 samples at the
same time. It has 4 data output lines (SDOA-D) that each contain the
data of one of the 4 channels. So it requires a SPI controller with 4
separate deserializers in order to receive all of the information at the
same time.

This should also work for the use case in [1] as well. (Some of the
patches in this series were already submitted there). In that case the
SPI controller is used kind of like it is two separate SPI controllers,
each with its own chip select, clock, and data lines.

[1]: https://lore.kernel.org/linux-spi/20250616220054.3968946-1-sean.anderson@linux.dev/

The DT bindings are a fairly straight-forward mapping of which pins on
the peripheral are connected to which pins on the controller. The SPI
core code parses this and makes the information available to drivers.
When a peripheral driver sees that multiple data lanes are wired up, it
can chose to use them when sending messages.

The SPI message API is a bit higher-level than just specifying the
number of data lines for a SPI transfer though. I did some research on
other SPI controllers that have this feature. They tend to be the kind
meant for connecting to two flash memory chips at the same time but can
be used more generically as well. They generally have the option to
either use one lane at a time (Sean's use case), or can mirror the same
data on multiple lanes (no users of this yet) or can perform striping
of a single data FIFO/DMA stream to/from the two lanes (our use case).

For now, the API assumes that if you want to do mirror/striping, then
you want to use all available data lanes. Otherwise, it just uses the
first data lane for "normal" SPI transfers.

14 files changed, 641 insertions, 27 deletions

diff --git a/Documentation/devicetree/bindings/display/panel/sitronix,st7789v.yaml b/Documentation/devicetree/bindings/display/panel/sitronix,st7789v.yaml
index 0ce2ea13583d..c35d4f2ab9a4 100644
--- a/Documentation/devicetree/bindings/display/panel/sitronix,st7789v.yaml
+++ b/Documentation/devicetree/bindings/display/panel/sitronix,st7789v.yaml
@@ -34,8 +34,9 @@ properties:
   spi-cpol: true
 
   spi-rx-bus-width:
-    minimum: 0
-    maximum: 1
+    items:
+      minimum: 0
+      maximum: 1
 
   dc-gpios:
     maxItems: 1
diff --git a/Documentation/devicetree/bindings/iio/adc/adi,ad4030.yaml b/Documentation/devicetree/bindings/iio/adc/adi,ad4030.yaml
index 54e7349317b7..e22d518135f2 100644
--- a/Documentation/devicetree/bindings/iio/adc/adi,ad4030.yaml
+++ b/Documentation/devicetree/bindings/iio/adc/adi,ad4030.yaml
@@ -37,7 +37,15 @@ properties:
     maximum: 102040816
 
   spi-rx-bus-width:
-    enum: [1, 2, 4]
+    maxItems: 2
+    # all lanes must have the same width
+    oneOf:
+      - contains:
+          const: 1
+      - contains:
+          const: 2
+      - contains:
+          const: 4
 
   vdd-5v-supply: true
   vdd-1v8-supply: true
@@ -88,6 +96,18 @@ oneOf:
 
 unevaluatedProperties: false
 
+allOf:
+  - if:
+      properties:
+        compatible:
+          enum:
+            - adi,ad4030-24
+            - adi,ad4032-24
+    then:
+      properties:
+        spi-rx-bus-width:
+          maxItems: 1
+
 examples:
   - |
     #include <dt-bindings/gpio/gpio.h>
@@ -108,3 +128,23 @@ examples:
             reset-gpios = <&gpio0 1 GPIO_ACTIVE_LOW>;
         };
     };
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+
+    spi {
+        #address-cells = <1>;
+        #size-cells = <0>;
+
+        adc@0 {
+            compatible = "adi,ad4630-24";
+            reg = <0>;
+            spi-max-frequency = <80000000>;
+            spi-rx-bus-width = <4>, <4>;
+            vdd-5v-supply = <&supply_5V>;
+            vdd-1v8-supply = <&supply_1_8V>;
+            vio-supply = <&supply_1_8V>;
+            ref-supply = <&supply_5V>;
+            cnv-gpios = <&gpio0 0 GPIO_ACTIVE_HIGH>;
+            reset-gpios = <&gpio0 1 GPIO_ACTIVE_LOW>;
+        };
+    };
diff --git a/Documentation/devicetree/bindings/iio/adc/adi,ad4695.yaml b/Documentation/devicetree/bindings/iio/adc/adi,ad4695.yaml
index cbde7a0505d2..ae8d0b5f328b 100644
--- a/Documentation/devicetree/bindings/iio/adc/adi,ad4695.yaml
+++ b/Documentation/devicetree/bindings/iio/adc/adi,ad4695.yaml
@@ -38,8 +38,9 @@ properties:
   spi-cpha: true
 
   spi-rx-bus-width:
-    minimum: 1
-    maximum: 4
+    items:
+      minimum: 1
+      maximum: 4
 
   avdd-supply:
     description: Analog power supply.
diff --git a/Documentation/devicetree/bindings/spi/adi,axi-spi-engine.yaml b/Documentation/devicetree/bindings/spi/adi,axi-spi-engine.yaml
index 4b3828eda6cb..0f2448371f17 100644
--- a/Documentation/devicetree/bindings/spi/adi,axi-spi-engine.yaml
+++ b/Documentation/devicetree/bindings/spi/adi,axi-spi-engine.yaml
@@ -70,6 +70,21 @@ required:
 
 unevaluatedProperties: false
 
+patternProperties:
+  "^.*@[0-9a-f]+":
+    type: object
+
+    properties:
+      spi-rx-bus-width:
+        maxItems: 8
+        items:
+          enum: [0, 1]
+
+      spi-tx-bus-width:
+        maxItems: 8
+        items:
+          enum: [0, 1]
+
 examples:
   - |
     spi@44a00000 {
diff --git a/Documentation/devicetree/bindings/spi/allwinner,sun4i-a10-spi.yaml b/Documentation/devicetree/bindings/spi/allwinner,sun4i-a10-spi.yaml
index e1ab3f523ad6..a34e6471dbe8 100644
--- a/Documentation/devicetree/bindings/spi/allwinner,sun4i-a10-spi.yaml
+++ b/Documentation/devicetree/bindings/spi/allwinner,sun4i-a10-spi.yaml
@@ -55,10 +55,12 @@ patternProperties:
           maximum: 4
 
       spi-rx-bus-width:
-        const: 1
+        items:
+          - const: 1
 
       spi-tx-bus-width:
-        const: 1
+        items:
+          - const: 1
 
 required:
   - compatible
diff --git a/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml b/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml
index 1b91d1566c95..a6067030c5ed 100644
--- a/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml
+++ b/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml
@@ -81,10 +81,12 @@ patternProperties:
           maximum: 4
 
       spi-rx-bus-width:
-        const: 1
+        items:
+          - const: 1
 
       spi-tx-bus-width:
-        const: 1
+        items:
+          - const: 1
 
 required:
   - compatible
diff --git a/Documentation/devicetree/bindings/spi/andestech,ae350-spi.yaml b/Documentation/devicetree/bindings/spi/andestech,ae350-spi.yaml
index 78093468dd5e..8e441742cee6 100644
--- a/Documentation/devicetree/bindings/spi/andestech,ae350-spi.yaml
+++ b/Documentation/devicetree/bindings/spi/andestech,ae350-spi.yaml
@@ -45,10 +45,12 @@ patternProperties:
 
     properties:
       spi-rx-bus-width:
-        enum: [1, 4]
+        items:
+          - enum: [1, 4]
 
       spi-tx-bus-width:
-        enum: [1, 4]
+        items:
+          - enum: [1, 4]
 
 allOf:
   - $ref: spi-controller.yaml#
diff --git a/Documentation/devicetree/bindings/spi/nvidia,tegra210-quad.yaml b/Documentation/devicetree/bindings/spi/nvidia,tegra210-quad.yaml
index 8b3640280559..909c204b8adf 100644
--- a/Documentation/devicetree/bindings/spi/nvidia,tegra210-quad.yaml
+++ b/Documentation/devicetree/bindings/spi/nvidia,tegra210-quad.yaml
@@ -54,10 +54,12 @@ patternProperties:
 
     properties:
       spi-rx-bus-width:
-        enum: [1, 2, 4]
+        items:
+          - enum: [1, 2, 4]
 
       spi-tx-bus-width:
-        enum: [1, 2, 4]
+        items:
+          - enum: [1, 2, 4]
 
 required:
   - compatible
diff --git a/Documentation/devicetree/bindings/spi/spi-peripheral-props.yaml b/Documentation/devicetree/bindings/spi/spi-peripheral-props.yaml
index 8b6e8fc009db..880a9f624566 100644
--- a/Documentation/devicetree/bindings/spi/spi-peripheral-props.yaml
+++ b/Documentation/devicetree/bindings/spi/spi-peripheral-props.yaml
@@ -64,9 +64,23 @@ properties:
     description:
       Bus width to the SPI bus used for read transfers.
       If 0 is provided, then no RX will be possible on this device.
-    $ref: /schemas/types.yaml#/definitions/uint32
-    enum: [0, 1, 2, 4, 8]
-    default: 1
+
+      Some SPI peripherals and controllers may have multiple data lanes for
+      receiving two or more words at the same time. If this is the case, each
+      index in the array represents the lane on both the SPI peripheral and
+      controller. Additional mapping properties may be needed if a lane is
+      skipped on either side.
+    $ref: /schemas/types.yaml#/definitions/uint32-array
+    items:
+      enum: [0, 1, 2, 4, 8]
+    default: [1]
+
+  spi-rx-lane-map:
+    description: Mapping of peripheral SDO lanes to controller SDI lanes.
+      Each index in the array represents a peripheral SDO lane, and the value
+      at that index represents the corresponding controller SDI lane.
+    $ref: /schemas/types.yaml#/definitions/uint32-array
+    default: [0, 1, 2, 3, 4, 5, 6, 7]
 
   spi-rx-delay-us:
     description:
@@ -81,9 +95,23 @@ properties:
     description:
       Bus width to the SPI bus used for write transfers.
       If 0 is provided, then no TX will be possible on this device.
-    $ref: /schemas/types.yaml#/definitions/uint32
-    enum: [0, 1, 2, 4, 8]
-    default: 1
+
+      Some SPI peripherals and controllers may have multiple data lanes for
+      transmitting two or more words at the same time. If this is the case, each
+      index in the array represents the lane on both the SPI peripheral and
+      controller. Additional mapping properties may be needed if a lane is
+      skipped on either side.
+    $ref: /schemas/types.yaml#/definitions/uint32-array
+    items:
+      enum: [0, 1, 2, 4, 8]
+    default: [1]
+
+  spi-tx-lane-map:
+    description: Mapping of peripheral SDI lanes to controller SDO lanes.
+      Each index in the array represents a peripheral SDI lane, and the value
+      at that index represents the corresponding controller SDO lane.
+    $ref: /schemas/types.yaml#/definitions/uint32-array
+    default: [0, 1, 2, 3, 4, 5, 6, 7]
 
   spi-tx-delay-us:
     description:
diff --git a/Documentation/spi/index.rst b/Documentation/spi/index.rst
index 824ce42ed4f0..2c89b1ee39e2 100644
--- a/Documentation/spi/index.rst
+++ b/Documentation/spi/index.rst
@@ -9,6 +9,7 @@ Serial Peripheral Interface (SPI)
 
    spi-summary
    spidev
+   multiple-data-lanes
    butterfly
    spi-lm70llp
    spi-sc18is602
diff --git a/Documentation/spi/multiple-data-lanes.rst b/Documentation/spi/multiple-data-lanes.rst
new file mode 100644
index 000000000000..69cb532d052f
--- /dev/null
+++ b/Documentation/spi/multiple-data-lanes.rst
@@ -0,0 +1,217 @@
+====================================
+SPI devices with multiple data lanes
+====================================
+
+Some specialized SPI controllers and peripherals support multiple data lanes
+that allow reading more than one word at a time in parallel. This is different
+from dual/quad/octal SPI where multiple bits of a single word are transferred
+simultaneously.
+
+For example, controllers that support parallel flash memories have this feature
+as do some simultaneous-sampling ADCs where each channel has its own data lane.
+
+---------------------
+Describing the wiring
+---------------------
+
+The ``spi-tx-bus-width`` and ``spi-rx-bus-width`` properties in the devicetree
+are used to describe how many data lanes are connected between the controller
+and how wide each lane is. The number of items in the array indicates how many
+lanes there are, and the value of each item indicates how many bits wide that
+lane is.
+
+For example, a dual-simultaneous-sampling ADC with two 4-bit lanes might be
+wired up like this::
+
+    +--------------+    +----------+
+    | SPI          |    | AD4630   |
+    | Controller   |    | ADC      |
+    |              |    |          |
+    |          CS0 |--->| CS       |
+    |          SCK |--->| SCK      |
+    |          SDO |--->| SDI      |
+    |              |    |          |
+    |        SDIA0 |<---| SDOA0    |
+    |        SDIA1 |<---| SDOA1    |
+    |        SDIA2 |<---| SDOA2    |
+    |        SDIA3 |<---| SDOA3    |
+    |              |    |          |
+    |        SDIB0 |<---| SDOB0    |
+    |        SDIB1 |<---| SDOB1    |
+    |        SDIB2 |<---| SDOB2    |
+    |        SDIB3 |<---| SDOB3    |
+    |              |    |          |
+    +--------------+    +----------+
+
+It is described in a devicetree like this::
+
+    spi {
+        compatible = "my,spi-controller";
+
+        ...
+
+        adc@0 {
+            compatible = "adi,ad4630";
+            reg = <0>;
+            ...
+            spi-rx-bus-width = <4>, <4>; /* 2 lanes of 4 bits each */
+            ...
+        };
+    };
+
+In most cases, lanes will be wired up symmetrically (A to A, B to B, etc). If
+this isn't the case, extra ``spi-rx-lane-map`` and ``spi-tx-lane-map``
+properties are needed to provide a mapping between controller lanes and the
+physical lane wires.
+
+Here is an example where a multi-lane SPI controller has each lane wired to
+separate single-lane peripherals::
+
+    +--------------+    +----------+
+    | SPI          |    | Thing 1  |
+    | Controller   |    |          |
+    |              |    |          |
+    |          CS0 |--->| CS       |
+    |         SDO0 |--->| SDI      |
+    |         SDI0 |<---| SDO      |
+    |        SCLK0 |--->| SCLK     |
+    |              |    |          |
+    |              |    +----------+
+    |              |
+    |              |    +----------+
+    |              |    | Thing 2  |
+    |              |    |          |
+    |          CS1 |--->| CS       |
+    |         SDO1 |--->| SDI      |
+    |         SDI1 |<---| SDO      |
+    |        SCLK1 |--->| SCLK     |
+    |              |    |          |
+    +--------------+    +----------+
+
+This is described in a devicetree like this::
+
+    spi {
+        compatible = "my,spi-controller";
+
+        ...
+
+        thing1@0 {
+            compatible = "my,thing1";
+            reg = <0>;
+            ...
+        };
+
+        thing2@1 {
+            compatible = "my,thing2";
+            reg = <1>;
+            ...
+            spi-tx-lane-map = <1>; /* lane 0 is not used, lane 1 is used for tx wire */
+            spi-rx-lane-map = <1>; /* lane 0 is not used, lane 1 is used for rx wire */
+            ...
+        };
+    };
+
+
+The default values of ``spi-rx-bus-width`` and ``spi-tx-bus-width`` are ``<1>``,
+so these properties can still be omitted even when ``spi-rx-lane-map`` and
+``spi-tx-lane-map`` are used.
+
+----------------------------
+Usage in a peripheral driver
+----------------------------
+
+These types of SPI controllers generally do not support arbitrary use of the
+multiple lanes. Instead, they operate in one of a few defined modes. Peripheral
+drivers should set the :c:type:`struct spi_transfer.multi_lane_mode <spi_transfer>`
+field to indicate which mode they want to use for a given transfer.
+
+The possible values for this field have the following semantics:
+
+- :c:macro:`SPI_MULTI_BUS_MODE_SINGLE`: Only use the first lane. Other lanes are
+    ignored. This means that it is operating just like a conventional SPI
+    peripheral. This is the default, so it does not need to be explicitly set.
+
+    Example::
+
+        tx_buf[0] = 0x88;
+
+        struct spi_transfer xfer = {
+            .tx_buf = tx_buf,
+            .len = 1,
+        };
+
+        spi_sync_transfer(spi, &xfer, 1);
+
+    Assuming the controller is sending the MSB first, the sequence of bits
+    sent over the tx wire would be (right-most bit is sent first)::
+
+        controller    > data bits >     peripheral
+        ----------   ----------------   ----------
+            SDO 0    0-0-0-1-0-0-0-1    SDI 0
+
+- :c:macro:`SPI_MULTI_BUS_MODE_MIRROR`: Send a single data word over all of the
+    lanes at the same time. This only makes sense for writes and not
+    for reads.
+
+    Example::
+
+        tx_buf[0] = 0x88;
+
+        struct spi_transfer xfer = {
+            .tx_buf = tx_buf,
+            .len = 1,
+            .multi_lane_mode = SPI_MULTI_BUS_MODE_MIRROR,
+        };
+
+        spi_sync_transfer(spi, &xfer, 1);
+
+    The data is mirrored on each tx wire::
+
+        controller    > data bits >     peripheral
+        ----------   ----------------   ----------
+            SDO 0    0-0-0-1-0-0-0-1    SDI 0
+            SDO 1    0-0-0-1-0-0-0-1    SDI 1
+
+- :c:macro:`SPI_MULTI_BUS_MODE_STRIPE`: Send or receive two different data words
+    at the same time, one on each lane. This means that the buffer needs to be
+    sized to hold data for all lanes. Data is interleaved in the buffer, with
+    the first word corresponding to lane 0, the second to lane 1, and so on.
+    Once the last lane is used, the next word in the buffer corresponds to lane
+    0 again. Accordingly, the buffer size must be a multiple of the number of
+    lanes. This mode works for both reads and writes.
+
+    Example::
+
+        struct spi_transfer xfer = {
+            .rx_buf = rx_buf,
+            .len = 2,
+            .multi_lane_mode = SPI_MULTI_BUS_MODE_STRIPE,
+        };
+
+        spi_sync_transfer(spi, &xfer, 1);
+
+    Each rx wire has a different data word sent simultaneously::
+
+        controller    < data bits <     peripheral
+        ----------   ----------------   ----------
+            SDI 0    0-0-0-1-0-0-0-1    SDO 0
+            SDI 1    1-0-0-0-1-0-0-0    SDO 1
+
+    After the transfer, ``rx_buf[0] == 0x11`` (word from SDO 0) and
+    ``rx_buf[1] == 0x88`` (word from SDO 1).
+
+
+-----------------------------
+SPI controller driver support
+-----------------------------
+
+To support multiple data lanes, SPI controller drivers need to set
+:c:type:`struct spi_controller.num_data_lanes <spi_controller>` to a value
+greater than 1.
+
+Then the part of the driver that handles SPI transfers needs to check the
+:c:type:`struct spi_transfer.multi_lane_mode <spi_transfer>` field and implement
+the appropriate behavior for each supported mode and return an error for
+unsupported modes.
+
+The core SPI code should handle the rest.
diff --git a/drivers/spi/spi-axi-spi-engine.c b/drivers/spi/spi-axi-spi-engine.c
index 91805eae9263..c75e8da049f7 100644
--- a/drivers/spi/spi-axi-spi-engine.c
+++ b/drivers/spi/spi-axi-spi-engine.c
@@ -23,6 +23,9 @@
 #include <linux/spi/spi.h>
 #include <trace/events/spi.h>
 
+#define SPI_ENGINE_REG_DATA_WIDTH		0x0C
+#define   SPI_ENGINE_REG_DATA_WIDTH_NUM_OF_SDIO_MASK	GENMASK(23, 16)
+#define   SPI_ENGINE_REG_DATA_WIDTH_MASK		GENMASK(15, 0)
 #define SPI_ENGINE_REG_OFFLOAD_MEM_ADDR_WIDTH	0x10
 #define SPI_ENGINE_REG_RESET			0x40
 
@@ -75,6 +78,8 @@
 #define SPI_ENGINE_CMD_REG_CLK_DIV		0x0
 #define SPI_ENGINE_CMD_REG_CONFIG		0x1
 #define SPI_ENGINE_CMD_REG_XFER_BITS		0x2
+#define SPI_ENGINE_CMD_REG_SDI_MASK		0x3
+#define SPI_ENGINE_CMD_REG_SDO_MASK		0x4
 
 #define SPI_ENGINE_MISC_SYNC			0x0
 #define SPI_ENGINE_MISC_SLEEP			0x1
@@ -105,6 +110,10 @@
 #define SPI_ENGINE_OFFLOAD_CMD_FIFO_SIZE	16
 #define SPI_ENGINE_OFFLOAD_SDO_FIFO_SIZE	16
 
+/* Extending SPI_MULTI_LANE_MODE values for optimizing messages. */
+#define SPI_ENGINE_MULTI_BUS_MODE_UNKNOWN	-1
+#define SPI_ENGINE_MULTI_BUS_MODE_CONFLICTING	-2
+
 struct spi_engine_program {
 	unsigned int length;
 	uint16_t instructions[] __counted_by(length);
@@ -142,6 +151,11 @@ struct spi_engine_offload {
 	unsigned long flags;
 	unsigned int offload_num;
 	unsigned int spi_mode_config;
+	unsigned int multi_lane_mode;
+	u8 rx_primary_lane_mask;
+	u8 tx_primary_lane_mask;
+	u8 rx_all_lanes_mask;
+	u8 tx_all_lanes_mask;
 	u8 bits_per_word;
 };
 
@@ -165,6 +179,25 @@ struct spi_engine {
 	bool offload_requires_sync;
 };
 
+static void spi_engine_primary_lane_flag(struct spi_device *spi,
+					 u8 *rx_lane_flags, u8 *tx_lane_flags)
+{
+	*rx_lane_flags = BIT(spi->rx_lane_map[0]);
+	*tx_lane_flags = BIT(spi->tx_lane_map[0]);
+}
+
+static void spi_engine_all_lanes_flags(struct spi_device *spi,
+				       u8 *rx_lane_flags, u8 *tx_lane_flags)
+{
+	int i;
+
+	for (i = 0; i < spi->num_rx_lanes; i++)
+		*rx_lane_flags |= BIT(spi->rx_lane_map[i]);
+
+	for (i = 0; i < spi->num_tx_lanes; i++)
+		*tx_lane_flags |= BIT(spi->tx_lane_map[i]);
+}
+
 static void spi_engine_program_add_cmd(struct spi_engine_program *p,
 	bool dry, uint16_t cmd)
 {
@@ -193,7 +226,7 @@ static unsigned int spi_engine_get_config(struct spi_device *spi)
 }
 
 static void spi_engine_gen_xfer(struct spi_engine_program *p, bool dry,
-	struct spi_transfer *xfer)
+				struct spi_transfer *xfer, u32 num_lanes)
 {
 	unsigned int len;
 
@@ -204,6 +237,9 @@ static void spi_engine_gen_xfer(struct spi_engine_program *p, bool dry,
 	else
 		len = xfer->len / 4;
 
+	if (xfer->multi_lane_mode == SPI_MULTI_LANE_MODE_STRIPE)
+		len /= num_lanes;
+
 	while (len) {
 		unsigned int n = min(len, 256U);
 		unsigned int flags = 0;
@@ -269,6 +305,7 @@ static int spi_engine_precompile_message(struct spi_message *msg)
 {
 	unsigned int clk_div, max_hz = msg->spi->controller->max_speed_hz;
 	struct spi_transfer *xfer;
+	int multi_lane_mode = SPI_ENGINE_MULTI_BUS_MODE_UNKNOWN;
 	u8 min_bits_per_word = U8_MAX;
 	u8 max_bits_per_word = 0;
 
@@ -284,6 +321,24 @@ static int spi_engine_precompile_message(struct spi_message *msg)
 			min_bits_per_word = min(min_bits_per_word, xfer->bits_per_word);
 			max_bits_per_word = max(max_bits_per_word, xfer->bits_per_word);
 		}
+
+		if (xfer->rx_buf || xfer->offload_flags & SPI_OFFLOAD_XFER_RX_STREAM ||
+		    xfer->tx_buf || xfer->offload_flags & SPI_OFFLOAD_XFER_TX_STREAM) {
+			switch (xfer->multi_lane_mode) {
+			case SPI_MULTI_LANE_MODE_SINGLE:
+			case SPI_MULTI_LANE_MODE_STRIPE:
+				break;
+			default:
+				/* Other modes, like mirror not supported */
+				return -EINVAL;
+			}
+
+			/* If all xfers have the same multi-lane mode, we can optimize. */
+			if (multi_lane_mode == SPI_ENGINE_MULTI_BUS_MODE_UNKNOWN)
+				multi_lane_mode = xfer->multi_lane_mode;
+			else if (multi_lane_mode != xfer->multi_lane_mode)
+				multi_lane_mode = SPI_ENGINE_MULTI_BUS_MODE_CONFLICTING;
+		}
 	}
 
 	/*
@@ -297,6 +352,14 @@ static int spi_engine_precompile_message(struct spi_message *msg)
 			priv->bits_per_word = min_bits_per_word;
 		else
 			priv->bits_per_word = 0;
+
+		priv->multi_lane_mode = multi_lane_mode;
+		spi_engine_primary_lane_flag(msg->spi,
+					     &priv->rx_primary_lane_mask,
+					     &priv->tx_primary_lane_mask);
+		spi_engine_all_lanes_flags(msg->spi,
+					   &priv->rx_all_lanes_mask,
+					   &priv->tx_all_lanes_mask);
 	}
 
 	return 0;
@@ -310,6 +373,7 @@ static void spi_engine_compile_message(struct spi_message *msg, bool dry,
 	struct spi_engine_offload *priv;
 	struct spi_transfer *xfer;
 	int clk_div, new_clk_div, inst_ns;
+	int prev_multi_lane_mode = SPI_MULTI_LANE_MODE_SINGLE;
 	bool keep_cs = false;
 	u8 bits_per_word = 0;
 
@@ -334,6 +398,7 @@ static void spi_engine_compile_message(struct spi_message *msg, bool dry,
 		 * in the same way.
 		 */
 		bits_per_word = priv->bits_per_word;
+		prev_multi_lane_mode = priv->multi_lane_mode;
 	} else {
 		spi_engine_program_add_cmd(p, dry,
 			SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_CONFIG,
@@ -344,6 +409,28 @@ static void spi_engine_compile_message(struct spi_message *msg, bool dry,
 	spi_engine_gen_cs(p, dry, spi, !xfer->cs_off);
 
 	list_for_each_entry(xfer, &msg->transfers, transfer_list) {
+		if (xfer->rx_buf || xfer->offload_flags & SPI_OFFLOAD_XFER_RX_STREAM ||
+		    xfer->tx_buf || xfer->offload_flags & SPI_OFFLOAD_XFER_TX_STREAM) {
+			if (xfer->multi_lane_mode != prev_multi_lane_mode) {
+				u8 tx_lane_flags, rx_lane_flags;
+
+				if (xfer->multi_lane_mode == SPI_MULTI_LANE_MODE_STRIPE)
+					spi_engine_all_lanes_flags(spi, &rx_lane_flags,
+								   &tx_lane_flags);
+				else
+					spi_engine_primary_lane_flag(spi, &rx_lane_flags,
+								     &tx_lane_flags);
+
+				spi_engine_program_add_cmd(p, dry,
+					SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDI_MASK,
+							     rx_lane_flags));
+				spi_engine_program_add_cmd(p, dry,
+					SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDO_MASK,
+							     tx_lane_flags));
+			}
+			prev_multi_lane_mode = xfer->multi_lane_mode;
+		}
+
 		new_clk_div = host->max_speed_hz / xfer->effective_speed_hz;
 		if (new_clk_div != clk_div) {
 			clk_div = new_clk_div;
@@ -360,7 +447,7 @@ static void spi_engine_compile_message(struct spi_message *msg, bool dry,
 					bits_per_word));
 		}
 
-		spi_engine_gen_xfer(p, dry, xfer);
+		spi_engine_gen_xfer(p, dry, xfer, spi->num_rx_lanes);
 		spi_engine_gen_sleep(p, dry, spi_delay_to_ns(&xfer->delay, xfer),
 				     inst_ns, xfer->effective_speed_hz);
 
@@ -394,6 +481,19 @@ static void spi_engine_compile_message(struct spi_message *msg, bool dry,
 	if (clk_div != 1)
 		spi_engine_program_add_cmd(p, dry,
 			SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_CLK_DIV, 0));
+
+	/* Restore single lane mode unless offload disable will restore it later. */
+	if (prev_multi_lane_mode == SPI_MULTI_LANE_MODE_STRIPE &&
+	    (!msg->offload || priv->multi_lane_mode != SPI_MULTI_LANE_MODE_STRIPE)) {
+		u8 rx_lane_flags, tx_lane_flags;
+
+		spi_engine_primary_lane_flag(spi, &rx_lane_flags, &tx_lane_flags);
+
+		spi_engine_program_add_cmd(p, dry,
+			SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDI_MASK, rx_lane_flags));
+		spi_engine_program_add_cmd(p, dry,
+			SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDO_MASK, tx_lane_flags));
+	}
 }
 
 static void spi_engine_xfer_next(struct spi_message *msg,
@@ -799,6 +899,19 @@ static int spi_engine_setup(struct spi_device *device)
 	writel_relaxed(SPI_ENGINE_CMD_CS_INV(spi_engine->cs_inv),
 		       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
 
+	if (host->num_data_lanes > 1) {
+		u8 rx_lane_flags, tx_lane_flags;
+
+		spi_engine_primary_lane_flag(device, &rx_lane_flags, &tx_lane_flags);
+
+		writel_relaxed(SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDI_MASK,
+						    rx_lane_flags),
+			       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
+		writel_relaxed(SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDO_MASK,
+						    tx_lane_flags),
+			       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
+	}
+
 	/*
 	 * In addition to setting the flags, we have to do a CS assert command
 	 * to make the new setting actually take effect.
@@ -902,6 +1015,15 @@ static int spi_engine_trigger_enable(struct spi_offload *offload)
 						    priv->bits_per_word),
 			       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
 
+	if (priv->multi_lane_mode == SPI_MULTI_LANE_MODE_STRIPE) {
+		writel_relaxed(SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDI_MASK,
+						    priv->rx_all_lanes_mask),
+			       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
+		writel_relaxed(SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDO_MASK,
+						    priv->tx_all_lanes_mask),
+			       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
+	}
+
 	writel_relaxed(SPI_ENGINE_CMD_SYNC(1),
 		spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
 
@@ -929,6 +1051,16 @@ static void spi_engine_trigger_disable(struct spi_offload *offload)
 	reg &= ~SPI_ENGINE_OFFLOAD_CTRL_ENABLE;
 	writel_relaxed(reg, spi_engine->base +
 			    SPI_ENGINE_REG_OFFLOAD_CTRL(priv->offload_num));
+
+	/* Restore single-lane mode. */
+	if (priv->multi_lane_mode == SPI_MULTI_LANE_MODE_STRIPE) {
+		writel_relaxed(SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDI_MASK,
+						    priv->rx_primary_lane_mask),
+			       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
+		writel_relaxed(SPI_ENGINE_CMD_WRITE(SPI_ENGINE_CMD_REG_SDO_MASK,
+						    priv->tx_primary_lane_mask),
+			       spi_engine->base + SPI_ENGINE_REG_CMD_FIFO);
+	}
 }
 
 static struct dma_chan
@@ -973,7 +1105,7 @@ static int spi_engine_probe(struct platform_device *pdev)
 {
 	struct spi_engine *spi_engine;
 	struct spi_controller *host;
-	unsigned int version;
+	unsigned int version, data_width_reg_val;
 	int irq, ret;
 
 	irq = platform_get_irq(pdev, 0);
@@ -1042,7 +1174,7 @@ static int spi_engine_probe(struct platform_device *pdev)
 		return PTR_ERR(spi_engine->base);
 
 	version = readl(spi_engine->base + ADI_AXI_REG_VERSION);
-	if (ADI_AXI_PCORE_VER_MAJOR(version) != 1) {
+	if (ADI_AXI_PCORE_VER_MAJOR(version) > 2) {
 		dev_err(&pdev->dev, "Unsupported peripheral version %u.%u.%u\n",
 			ADI_AXI_PCORE_VER_MAJOR(version),
 			ADI_AXI_PCORE_VER_MINOR(version),
@@ -1050,6 +1182,8 @@ static int spi_engine_probe(struct platform_device *pdev)
 		return -ENODEV;
 	}
 
+	data_width_reg_val = readl(spi_engine->base + SPI_ENGINE_REG_DATA_WIDTH);
+
 	if (adi_axi_pcore_ver_gteq(version, 1, 1)) {
 		unsigned int sizes = readl(spi_engine->base +
 				SPI_ENGINE_REG_OFFLOAD_MEM_ADDR_WIDTH);
@@ -1096,6 +1230,9 @@ static int spi_engine_probe(struct platform_device *pdev)
 	}
 	if (adi_axi_pcore_ver_gteq(version, 1, 3))
 		host->mode_bits |= SPI_MOSI_IDLE_LOW | SPI_MOSI_IDLE_HIGH;
+	if (adi_axi_pcore_ver_gteq(version, 2, 0))
+		host->num_data_lanes = FIELD_GET(SPI_ENGINE_REG_DATA_WIDTH_NUM_OF_SDIO_MASK,
+						 data_width_reg_val);
 
 	if (host->max_speed_hz == 0)
 		return dev_err_probe(&pdev->dev, -EINVAL, "spi_clk rate is 0");
diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index b773c297f8b1..a24de5c98399 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -2354,8 +2354,8 @@ static void of_spi_parse_dt_cs_delay(struct device_node *nc,
 static int of_spi_parse_dt(struct spi_controller *ctlr, struct spi_device *spi,
 			   struct device_node *nc)
 {
-	u32 value, cs[SPI_DEVICE_CS_CNT_MAX];
-	int rc, idx;
+	u32 value, cs[SPI_DEVICE_CS_CNT_MAX], map[SPI_DEVICE_DATA_LANE_CNT_MAX];
+	int rc, idx, max_num_data_lanes;
 
 	/* Mode (clock phase/polarity/etc.) */
 	if (of_property_read_bool(nc, "spi-cpha"))
@@ -2370,7 +2370,65 @@ static int of_spi_parse_dt(struct spi_controller *ctlr, struct spi_device *spi,
 		spi->mode |= SPI_CS_HIGH;
 
 	/* Device DUAL/QUAD mode */
-	if (!of_property_read_u32(nc, "spi-tx-bus-width", &value)) {
+
+	rc = of_property_read_variable_u32_array(nc, "spi-tx-lane-map", map, 1,
+						 ARRAY_SIZE(map));
+	if (rc >= 0) {
+		max_num_data_lanes = rc;
+		for (idx = 0; idx < max_num_data_lanes; idx++)
+			spi->tx_lane_map[idx] = map[idx];
+	} else if (rc == -EINVAL) {
+		/* Default lane map is identity mapping. */
+		max_num_data_lanes = ARRAY_SIZE(spi->tx_lane_map);
+		for (idx = 0; idx < max_num_data_lanes; idx++)
+			spi->tx_lane_map[idx] = idx;
+	} else {
+		dev_err(&ctlr->dev,
+			"failed to read spi-tx-lane-map property: %d\n", rc);
+		return rc;
+	}
+
+	rc = of_property_count_u32_elems(nc, "spi-tx-bus-width");
+	if (rc < 0 && rc != -EINVAL) {
+		dev_err(&ctlr->dev,
+			"failed to read spi-tx-bus-width property: %d\n", rc);
+		return rc;
+	}
+	if (rc > max_num_data_lanes) {
+		dev_err(&ctlr->dev,
+			"spi-tx-bus-width has more elements (%d) than spi-tx-lane-map (%d)\n",
+			rc, max_num_data_lanes);
+		return -EINVAL;
+	}
+
+	if (rc == -EINVAL) {
+		/* Default when property is not present. */
+		spi->num_tx_lanes = 1;
+	} else {
+		u32 first_value;
+
+		spi->num_tx_lanes = rc;
+
+		for (idx = 0; idx < spi->num_tx_lanes; idx++) {
+			rc = of_property_read_u32_index(nc, "spi-tx-bus-width",
+							idx, &value);
+			if (rc)
+				return rc;
+
+			/*
+			 * For now, we only support all lanes having the same
+			 * width so we can keep using the existing mode flags.
+			 */
+			if (!idx)
+				first_value = value;
+			else if (first_value != value) {
+				dev_err(&ctlr->dev,
+					"spi-tx-bus-width has inconsistent values: first %d vs later %d\n",
+					first_value, value);
+				return -EINVAL;
+			}
+		}
+
 		switch (value) {
 		case 0:
 			spi->mode |= SPI_NO_TX;
@@ -2394,7 +2452,74 @@ static int of_spi_parse_dt(struct spi_controller *ctlr, struct spi_device *spi,
 		}
 	}
 
-	if (!of_property_read_u32(nc, "spi-rx-bus-width", &value)) {
+	for (idx = 0; idx < spi->num_tx_lanes; idx++) {
+		if (spi->tx_lane_map[idx] >= spi->controller->num_data_lanes) {
+			dev_err(&ctlr->dev,
+				"spi-tx-lane-map has invalid value %d (num_data_lanes=%d)\n",
+				spi->tx_lane_map[idx],
+				spi->controller->num_data_lanes);
+			return -EINVAL;
+		}
+	}
+
+	rc = of_property_read_variable_u32_array(nc, "spi-rx-lane-map", map, 1,
+						 ARRAY_SIZE(map));
+	if (rc >= 0) {
+		max_num_data_lanes = rc;
+		for (idx = 0; idx < max_num_data_lanes; idx++)
+			spi->rx_lane_map[idx] = map[idx];
+	} else if (rc == -EINVAL) {
+		/* Default lane map is identity mapping. */
+		max_num_data_lanes = ARRAY_SIZE(spi->rx_lane_map);
+		for (idx = 0; idx < max_num_data_lanes; idx++)
+			spi->rx_lane_map[idx] = idx;
+	} else {
+		dev_err(&ctlr->dev,
+			"failed to read spi-rx-lane-map property: %d\n", rc);
+		return rc;
+	}
+
+	rc = of_property_count_u32_elems(nc, "spi-rx-bus-width");
+	if (rc < 0 && rc != -EINVAL) {
+		dev_err(&ctlr->dev,
+			"failed to read spi-rx-bus-width property: %d\n", rc);
+		return rc;
+	}
+	if (rc > max_num_data_lanes) {
+		dev_err(&ctlr->dev,
+			"spi-rx-bus-width has more elements (%d) than spi-rx-lane-map (%d)\n",
+			rc, max_num_data_lanes);
+		return -EINVAL;
+	}
+
+	if (rc == -EINVAL) {
+		/* Default when property is not present. */
+		spi->num_rx_lanes = 1;
+	} else {
+		u32 first_value;
+
+		spi->num_rx_lanes = rc;
+
+		for (idx = 0; idx < spi->num_rx_lanes; idx++) {
+			rc = of_property_read_u32_index(nc, "spi-rx-bus-width",
+							idx, &value);
+			if (rc)
+				return rc;
+
+			/*
+			 * For now, we only support all lanes having the same
+			 * width so we can keep using the existing mode flags.
+			 */
+			if (!idx)
+				first_value = value;
+			else if (first_value != value) {
+				dev_err(&ctlr->dev,
+					"spi-rx-bus-width has inconsistent values: first %d vs later %d\n",
+					first_value, value);
+				return -EINVAL;
+			}
+		}
+
 		switch (value) {
 		case 0:
 			spi->mode |= SPI_NO_RX;
@@ -2418,6 +2543,16 @@ static int of_spi_parse_dt(struct spi_controller *ctlr, struct spi_device *spi,
 		}
 	}
 
+	for (idx = 0; idx < spi->num_rx_lanes; idx++) {
+		if (spi->rx_lane_map[idx] >= spi->controller->num_data_lanes) {
+			dev_err(&ctlr->dev,
+				"spi-rx-lane-map has invalid value %d (num_data_lanes=%d)\n",
+				spi->rx_lane_map[idx],
+				spi->controller->num_data_lanes);
+			return -EINVAL;
+		}
+	}
+
 	if (spi_controller_is_target(ctlr)) {
 		if (!of_node_name_eq(nc, "slave")) {
 			dev_err(&ctlr->dev, "%pOF is not called 'slave'\n",
@@ -3066,6 +3201,7 @@ struct spi_controller *__spi_alloc_controller(struct device *dev,
 	mutex_init(&ctlr->add_lock);
 	ctlr->bus_num = -1;
 	ctlr->num_chipselect = 1;
+	ctlr->num_data_lanes = 1;
 	ctlr->target = target;
 	if (IS_ENABLED(CONFIG_SPI_SLAVE) && target)
 		ctlr->dev.class = &spi_target_class;
diff --git a/include/linux/spi/spi.h b/include/linux/spi/spi.h
index cb2c2df31089..39681f7e063b 100644
--- a/include/linux/spi/spi.h
+++ b/include/linux/spi/spi.h
@@ -23,6 +23,9 @@
 /* Max no. of CS supported per spi device */
 #define SPI_DEVICE_CS_CNT_MAX 4
 
+/* Max no. of data lanes supported per spi device */
+#define SPI_DEVICE_DATA_LANE_CNT_MAX 8
+
 struct dma_chan;
 struct software_node;
 struct ptp_system_timestamp;
@@ -174,6 +177,10 @@ extern void spi_transfer_cs_change_delay_exec(struct spi_message *msg,
  * @cs_index_mask: Bit mask of the active chipselect(s) in the chipselect array
  * @cs_gpiod: Array of GPIO descriptors of the corresponding chipselect lines
  *	(optional, NULL when not using a GPIO line)
+ * @tx_lane_map: Map of peripheral lanes (index) to controller lanes (value).
+ * @num_tx_lanes: Number of transmit lanes wired up.
+ * @rx_lane_map: Map of peripheral lanes (index) to controller lanes (value).
+ * @num_rx_lanes: Number of receive lanes wired up.
  *
  * A @spi_device is used to interchange data between an SPI target device
  * (usually a discrete chip) and CPU memory.
@@ -242,6 +249,12 @@ struct spi_device {
 
 	struct gpio_desc	*cs_gpiod[SPI_DEVICE_CS_CNT_MAX];	/* Chip select gpio desc */
 
+	/* Multi-lane SPI controller support. */
+	u8			tx_lane_map[SPI_DEVICE_DATA_LANE_CNT_MAX];
+	u8			num_tx_lanes;
+	u8			rx_lane_map[SPI_DEVICE_DATA_LANE_CNT_MAX];
+	u8			num_rx_lanes;
+
 	/*
 	 * Likely need more hooks for more protocol options affecting how
 	 * the controller talks to each chip, like:
@@ -401,6 +414,7 @@ extern struct spi_device *spi_new_ancillary_device(struct spi_device *spi, u8 ch
  *	SPI targets, and are numbered from zero to num_chipselects.
  *	each target has a chipselect signal, but it's common that not
  *	every chipselect is connected to a target.
+ * @num_data_lanes: Number of data lanes supported by this controller. Default is 1.
  * @dma_alignment: SPI controller constraint on DMA buffers alignment.
  * @mode_bits: flags understood by this controller driver
  * @buswidth_override_bits: flags to override for this controller driver
@@ -576,6 +590,14 @@ struct spi_controller {
 	 */
 	u16			num_chipselect;
 
+	/*
+	 * Some specialized SPI controllers can have more than one physical
+	 * data lane interface per controller (each having it's own serializer).
+	 * This specifies the number of data lanes in that case. Other
+	 * controllers do not need to set this (defaults to 1).
+	 */
+	u16			num_data_lanes;
+
 	/* Some SPI controllers pose alignment requirements on DMAable
 	 * buffers; let protocol drivers know about these requirements.
 	 */
@@ -959,6 +981,8 @@ struct spi_res {
  *      (SPI_NBITS_SINGLE) is used.
  * @rx_nbits: number of bits used for reading. If 0 the default
  *      (SPI_NBITS_SINGLE) is used.
+ * @multi_lane_mode: How to serialize data on multiple lanes. One of the
+ *      SPI_MULTI_LANE_MODE_* values.
  * @len: size of rx and tx buffers (in bytes)
  * @speed_hz: Select a speed other than the device default for this
  *      transfer. If 0 the default (from @spi_device) is used.
@@ -1095,6 +1119,12 @@ struct spi_transfer {
 	unsigned	cs_change:1;
 	unsigned	tx_nbits:4;
 	unsigned	rx_nbits:4;
+
+#define SPI_MULTI_LANE_MODE_SINGLE	0 /* only use single lane */
+#define SPI_MULTI_LANE_MODE_STRIPE	1 /* one data word per lane */
+#define SPI_MULTI_LANE_MODE_MIRROR	2 /* same word sent on all lanes */
+	unsigned	multi_lane_mode: 2;
+
 	unsigned	timestamped:1;
 	bool		dtr_mode;
 #define	SPI_NBITS_SINGLE	0x01 /* 1-bit transfer */